A pepper harvesting device operating speed self-adaptive control system and method

Abstract

The application discloses a pepper harvesting equipment operation speed self-adaptive control system and method, and relates to the technical field of mechanical control. The system comprises the following steps: synchronously collecting the instantaneous value of the roller torque and the instantaneous value of the walking speed through fixed sampling intervals to reconstruct the phase space; the system comprises a phase trajectory generation module, which combines the instantaneous value of the torque and the walking speed to form an augmented phase space trajectory and calculates the spatial distance between each phase point; a recursive feature construction module forms a recursive matrix by judging the distance between the phase points; a recursive feature judgment module counts the distribution density of the recursive points in the diagonal direction, extracts the length distribution feature to calculate the recursive trend intensity; a speed control amount generation module judges the jamming state type according to the change rate of the recursive trend intensity, and generates a speed reduction adjustment amount according to a nonlinear compression function to optimize the operation speed, improve the harvesting efficiency and safety.

Description

Technical Field

[0001] This invention relates to the field of mechanical control technology, specifically to an adaptive control system and method for the operating speed of chili harvesting equipment. Background Technology

[0002] In agriculture, especially during the harvesting of crops such as chili peppers, the operating speed of harvesting equipment directly affects operational efficiency and crop quality. With the gradual promotion of agricultural mechanization, modern chili pepper harvesting equipment is typically equipped with efficient power systems and intelligent control systems. These devices need to adjust their operating speed in real time according to changes in the field environment, such as soil conditions, plant growth status, and operating resistance, to ensure smooth harvesting. However, existing technologies often employ fixed-speed or simple feedback control strategies, which cannot fully adapt to complex and changing operating environments.

[0003] Existing technologies mainly rely on field data collected by sensors to adjust equipment speed through simple PID control or empirical adjustments. While these methods can maintain equipment performance in static environments, they often fail to respond quickly to changing external conditions in dynamic environments, such as changes in chili plant density or mechanical failures. This can lead to problems such as insufficient power, blockages, or damage, severely impacting harvesting efficiency and crop integrity. Furthermore, existing technologies lack specificity in assessing the mechanical stability of the equipment and developing speed adjustment strategies, making it difficult to effectively prevent abnormal situations during operation.

[0004] Therefore, there is an urgent need to significantly improve the equipment's adaptability to complex environments, address the shortcomings of existing technologies that fail to respond to environmental changes in real time, improve chili harvesting efficiency, and reduce damage rates.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an adaptive control system and method for the operating speed of chili harvesting equipment, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An adaptive control system for the operating speed of a chili harvesting device, specifically comprising: The spatial phase point construction module is used to synchronously collect the instantaneous values ​​of the drum torque and the walking speed of the harvesting equipment during the chili harvesting operation at fixed sampling intervals, and to reconstruct the phase space of the torque instantaneous value sequence at multiple consecutive sampling times according to preset time delay embedding parameters. The phase trajectory generation module is used to combine the instantaneous torque value at any moment with the instantaneous torque value of the preceding time delay at that moment to jointly form the phase point corresponding to that moment. The current walking speed value is added as an additional dimension to each phase point, and the phase points are collected to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, the spatial distance between it and other phase points in the trajectory is calculated. The recursive feature construction module is used to determine whether two phase points constitute a recursive state based on the spatial distance between each phase point and a preset distance neighborhood threshold, and to count the phase points that constitute a recursive state into a recursive matrix with time as the row and column index. The recursive feature judgment module is used to statistically analyze the distribution density of recursive points along the diagonal direction in the recursive matrix, extract the length distribution features of recursive segments along the diagonal direction, and calculate the recursive trend strength reflecting the mechanical stability of the harvesting equipment based on the length distribution features. The speed control quantity generation module is used to compare the recursive trend intensity at the current moment with the recursive trend intensity at the previous sampling moment, calculate the rate of change, determine the type of blockage based on the rate of change and the current recursive trend intensity, and generate the speed reduction adjustment amount by combining the blockage type, the current recursive trend intensity, and the preset nonlinear compression function.

[0008] Furthermore, the time delay embedding parameters specifically include time delay and phase point embedding dimension; the specific method for determining time delay and phase point embedding dimension is as follows: design field chili harvesting experiments, collect instantaneous torque values ​​and instantaneous walking speed values ​​under different harvesting conditions, the harvesting conditions include unloaded uniform speed, normal feeding, soil impact, gradual blockage and recovery after manual intervention to reduce speed; Based on the experimental roller torque time series data, the mutual information method is used to calculate the mutual information value between any roller torque time series data and the roller torque after the corresponding time delay. The specific steps are as follows: set several fixed time delays, which are set based on the sampling interval; calculate the mutual information value of the roller torque time series data under different fixed time delays; analyze the changes of all mutual information values ​​under each fixed time delay; and when the mutual information value first drops to a preset information threshold, use the corresponding fixed time delay as the time delay for setting the phase point. Given a fixed time delay, the phase point embedding dimension is set to increase incrementally. The pseudo-nearest neighbor method is used to calculate the pseudo-nearest neighbor ratio for different embedding dimensions. If the pseudo-nearest neighbor ratio drops to a preset ratio threshold or below for the first time, and the decrease in the pseudo-nearest neighbor ratio does not exceed a preset amplitude threshold when the embedding dimension is subsequently increased, then the embedding dimension is taken as the fixed embedding dimension.

[0009] Furthermore, phase points are constructed based on the time delay embedding parameters, specifically using the following formula: In the formula, Let k be the phase point at the k-th sampling time. Let be the instantaneous value of the torque at the k-th sampling time. For the k-th The instantaneous value of torque at the sampling time. For a fixed time delay, To fix the embedding dimension, Let k be the instantaneous walking speed at the k-th sampling time; k is the index of the sampling time. For any two phase points in the augmented phase space trajectory, their Euclidean distance is calculated as the spatial distance. The specific logic behind this calculation is as follows: In the formula, Specifically, it represents the k-th phase point and the... Spatial distance between phase points Represents the kth phase point. One portion, Indicates the first The th phase point Each component, k and Let be the index of the phase point, and .

[0010] Furthermore, the specific method for obtaining the preset distance neighborhood threshold is as follows: for any phase point in the current time window, determine the three phase points that are spatially closest to it, and define the spatial distance between the phase point and the three closest phase points as three sets of nearest neighbor distances. Traverse all phase points in the current time window, summarize the three nearest neighbor distances of all phase points into a set, and use the product of the median of the set and the preset margin coefficient as the distance neighborhood threshold. The current time window is specifically defined as having the current time as the end point of the time window and determining the start point of the time window through a preset time width, thereby forming the current time window.

[0011] Furthermore, the specific method for constructing the recursive matrix is ​​as follows: The spatial distance between different phase points is compared with a distance neighborhood threshold. If the spatial distance between different phase points is not greater than the distance neighborhood threshold, then the two phase points constitute a recursive state. This is used to construct a binary recursive matrix, whose elements are defined as follows: In the formula, For the k-th row of the recursive matrix, the first... Column element values; The logic behind extracting the length distribution features of consecutive recursive segments along the diagonal is as follows: Diagonal structure analysis is performed on the constructed recursive matrix, and from any recursive matrix... The process begins by scanning along the main diagonal, identifying segments with consecutive element values ​​of 1 along this direction. The length of each segment is recorded, with a minimum interval of 1 between adjacent element values. Segments with a length of at least 2 are designated as target segments. The frequency of each target segment is recorded, and the recursive trend strength is calculated based on this frequency, using the following formula: In the formula, For the strength of the recursive trend, The length of the target segment. The maximum length of the target segment. This indicates the frequency of occurrence of a target segment of length c.

[0012] Furthermore, a preset dynamic threshold value is used to determine the blockage status. This dynamic threshold value is specifically calculated dynamically based on historical data, while also incorporating a forgetting factor to ensure that the threshold value fluctuates with the normal operating conditions of the system. The specific formula used is as follows: In the formula, Let be the dynamic boundary value for the t-th time window. The recursive trend strength is the value of the (t-1)th time window. Forgetting factor, This is the dynamic boundary value for the (t-1)th time window, where t is the index of the time window; The specific formula used to calculate the rate of change of the recursive trend strength is as follows: In the formula, Let be the rate of change of the recursive trend intensity in the t-th time window. Let be the recursive trend strength at the t-th time window; The current congestion status is determined based on the rate of change of recursive trend intensity combined with a dynamic boundary value. The specific logic is as follows: If both conditions are met Less than 0, and When the value is less than the current dynamic threshold, it is determined that the system is currently in a state of congestion. when Not less than 0 or If the value is not less than the current dynamic boundary value, it is determined that the current state is non-blocking.

[0013] Furthermore, in non-blocking conditions, no adjustment of the operating speed is required; For the congestion development state, a nonlinear compression function is constructed. The nonlinear compression function is specifically set based on the intensity of the recursive trend within the time window, and the specific formula is as follows: In the formula, Let be the deceleration adjustment amount for the t-th time window. The maximum single speed adjustment range is set. This is the sensitivity coefficient; Based on the deceleration adjustment amount in the t-th time window, a first-order low-pass smoothing is applied to obtain the final target speed. The specific formula used to calculate the target speed is as follows: In the formula, Let be the target speed for the t-th time window. Let be the average actual walking speed in the t-th time window. Let the target velocity be the velocity in the (t-1)th time window. For smoothing coefficients; If the congestion type is a congestion development state, the speed reduction adjustment is the calculated value obtained by substituting the recursive trend strength into a preset nonlinear compression function and performing first-order low-pass smoothing. If the congestion type is a non-congestion state, the speed reduction adjustment is zero.

[0014] The present invention also provides an adaptive control method for the operating speed of a chili harvesting device, wherein the adaptive control method for the operating speed of a chili harvesting device is used to control the aforementioned adaptive control system for the operating speed of a chili harvesting device, comprising: During chili harvesting operations, the instantaneous values ​​of the drum torque and travel speed of the harvesting equipment are collected synchronously at fixed sampling intervals. The sequence of instantaneous torque values ​​from multiple consecutive sampling times is then reconstructed in phase space according to preset time delay embedding parameters. The instantaneous torque value at any given moment is combined with the instantaneous torque value of the preceding time delay to form a phase point corresponding to that moment. The current walking speed value is added as an additional dimension to each phase point. The phase points are aggregated to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, the spatial distance between it and other phase points in the trajectory is calculated. Based on the spatial distance between each phase point, a preset distance neighborhood threshold is used to determine whether two phase points constitute a recursive state. The phase points that constitute a recursive state are counted as a recursive matrix with time as the row and column index. The distribution density of recursion points along the diagonal direction in the recursion matrix is ​​statistically analyzed, the length distribution characteristics of recursion segments along the diagonal direction are extracted, and the recursion trend strength reflecting the mechanical stability of the harvesting equipment is calculated based on the length distribution characteristics. The recursive trend strength at the current moment is compared with the recursive trend strength at the previous sampling moment to obtain the rate of change. The type of congestion is determined based on the rate of change and the current recursive trend strength. The deceleration adjustment amount is generated by combining the congestion type, the current recursive trend strength, and the preset nonlinear compression function.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention synchronously collects roller torque and travel speed data at fixed sampling intervals, and uses this as a basis to reconstruct the phase space, forming multi-dimensional phase points and calculating the spatial relationship between the phase points; it effectively reflects the real-time status of the equipment during operation, and also deeply explores the mechanical stability of the equipment during operation through the analysis of the recursive matrix; thus, the system can judge the working status of the equipment in real time, especially when facing external disturbances, and promptly identify possible blockages, thereby taking corresponding speed reduction adjustment measures. This invention also constructs an augmented phase space trajectory, which can characterize the dynamic behavior of the equipment in detail; calculates the spatial distance between phase points and determines the recursive state, enabling the system to dynamically adjust the operating speed of the equipment based on historical data and the current operating state; the adaptive control mechanism can effectively avoid fruit damage and equipment failure caused by improper speed, thus improving harvest quality; in the harvesting process of vulnerable crops such as chili peppers, the control system can prevent fruit damage caused by excessive torque at high speeds, ensuring the integrity of the harvest and its commercial value; Furthermore, the introduction of recursive trend strength into the system provides a new method for assessing mechanical stability. By statistically analyzing the distribution density of recursion points, the stability of the equipment can be quantified, enabling the system to not only adapt to the current operating environment but also continuously optimize its control strategy based on the equipment's operating history, thereby continuously improving harvesting efficiency and safety during long-term operation. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall system structure of the present invention; Figure 2 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0018] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0019] Example: Please see Figure 1 The present invention provides a technical solution: An adaptive control system for the operating speed of a chili harvesting device, specifically comprising: The spatial phase point construction module is used to synchronously collect the instantaneous values ​​of the drum torque and the walking speed of the harvesting equipment during chili harvesting operations at fixed sampling intervals, and to reconstruct the phase space of the instantaneous torque value sequence at multiple consecutive sampling times according to preset time delay embedding parameters.

[0020] The time delay embedding parameters specifically include time delay and phase point embedding dimension; the specific method for determining time delay and phase point embedding dimension is as follows: design field chili harvesting test, collect instantaneous torque value and instantaneous walking speed value under different harvesting conditions, the harvesting conditions include unloaded uniform speed, normal feeding, soil impact, gradual blockage and recovery after manual intervention to reduce speed; Specifically, this includes: selecting suitable field plots for chili cultivation, ensuring they are representative and can cover different soil types and growing conditions, such as soil moisture and hardness, ensuring the normal operation of chili harvesting equipment, and equipping it with high-precision sensors for real-time monitoring of roller torque and travel speed. The equipment travels at a constant speed when there are no chili peppers, and the instantaneous torque and travel speed values ​​are recorded under this condition. When the equipment is harvesting chili peppers normally, the instantaneous torque and travel speed values ​​are recorded. During the harvesting process, the equipment encounters clods of soil, causing instantaneous torque changes, and the instantaneous torque and travel speed values ​​are recorded under this condition. The equipment simulates a gradually increasing load during the harvesting period, causing the equipment torque to rise, and the instantaneous torque and travel speed values ​​are recorded during this process. During the harvesting process, the equipment is forced to decelerate due to operational intervention, and then returns to normal speed, and the instantaneous torque and travel speed values ​​are recorded. A fixed sampling interval is set based on the equipment performance and sensor response time to ensure that changes in instantaneous torque and walking speed values ​​can be captured.

[0021] Based on the experimental roller torque time series data, the mutual information method is used to calculate the mutual information value between any roller torque time series data and the roller torque after the corresponding time delay. The specific steps are as follows: set several fixed time delays, which are set based on the sampling interval; calculate the mutual information value of the roller torque time series data under different fixed time delays; analyze the changes of all mutual information values ​​under each fixed time delay; and when the mutual information value first drops to a preset information threshold, use the corresponding fixed time delay as the time delay for setting the phase point. Specifically, this includes: based on roller torque timing data and the corresponding time-delayed data The marginal probability distributions of each are statistically analyzed. Based on the marginal probability distributions, the mutual information values ​​under each fixed time delay are calculated according to the mutual information calculation formula, and their changing trends are analyzed. Given a fixed time delay, the phase point embedding dimension is set to increase incrementally. The pseudo-nearest neighbor method is used to calculate the pseudo-nearest neighbor ratio for different embedding dimensions. If the pseudo-nearest neighbor ratio drops to a preset ratio threshold or below for the first time, and the decrease in the pseudo-nearest neighbor ratio does not exceed a preset amplitude threshold when the embedding dimension is subsequently increased, then the embedding dimension is taken as the fixed embedding dimension.

[0022] The preset threshold ratio is used to determine the proportion of false nearest neighbors. It represents the critical value at which the proportion of false nearest neighbors needs to be reduced to judge whether the embedding dimension is reasonable. It is usually set to a small proportion value, with a common range of 5%-10%. That is, when the proportion of false nearest neighbors is reduced to within 5%-10%, the influence of false nearest neighbors is considered negligible. For systems with a large data sample size, a lower threshold, such as 5%, can be selected to improve accuracy; while for systems with a small sample size, the threshold can be increased to 10%-15% as appropriate.

[0023] A preset amplitude threshold is used to limit the rate of decrease in the proportion of false nearest neighbors as the embedding dimension gradually increases. Its purpose is to avoid excessively increasing the embedding dimension, ensuring that the chosen embedding dimension is both reasonable and computationally efficient. As the embedding dimension gradually increases, the rate of decrease in the proportion of false nearest neighbors should tend to plateau. If the rate of decrease in the proportion of false nearest neighbors is small, it indicates that further increasing the embedding dimension will not yield significant benefits. The preset amplitude threshold is generally set between 1% and 2%. For simple systems, the rate of decrease in the proportion of false nearest neighbors may plateau relatively quickly, so a smaller amplitude threshold, such as 1%, can be chosen. For complex systems, the threshold can be appropriately increased, such as 2% to 3%.

[0024] When the embedding dimension is insufficient, two seemingly close phase points in a low-dimensional reconstruction space may actually be spurious projections due to the insufficient dimension. Upgrading the dimension to... After dimensionality, if the distance between two points increases sharply, it is determined to be a false nearest neighbor; when the proportion of true nearest neighbors increases... When the proportion of false nearest neighbors increases and then stabilizes, and decreases to near zero, the corresponding... This is the minimum fully embedded dimension.

[0025] The phase trajectory generation module is used to combine the instantaneous torque value at any given moment with the instantaneous torque value of the preceding time delay to jointly form a phase point corresponding to that moment. The current walking speed value is added as an additional dimension to each phase point, and the phase points are aggregated to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, the spatial distance between it and other phase points in the trajectory is calculated.

[0026] The phase point is constructed based on the time delay embedding parameters, specifically using the following formula: In the formula, Let k be the phase point at the k-th sampling time. Let be the instantaneous value of the torque at the k-th sampling time. For the k-th The instantaneous value of torque at the sampling time. For a fixed time delay, To fix the embedding dimension, Let k be the instantaneous walking speed at the k-th sampling time; k is the index of the sampling time. It should be noted that, Indicates the first The phase point constructed at each sampling time is a high-dimensional vector containing the state information of the current time and several previous times, as well as the walking speed at the current time; Instantaneous torque value at the current moment Torque values ​​at previous times These are combined to effectively capture the dynamic characteristics of the system, reflect the trend of torque changes over time, and provide rich historical information for subsequent analysis; By introducing a fixed time delay This ensures that the phase points are constructed based on a unified time frame; all phase points The convergence of phase space trajectories provides a multidimensional perspective for subsequent analysis; the way the phase space is constructed fully reflects the relationship between different time points, facilitating subsequent analysis such as prediction, control, and pattern recognition.

[0027] For any two phase points in the augmented phase space trajectory, their Euclidean distance is calculated as the spatial distance. The specific logic behind this calculation is as follows: In the formula, Specifically, it represents the k-th phase point and the... Spatial distance between phase points Represents the kth phase point. One portion, Indicates the first The th phase point Each component, k and Let be the index of the phase point, and .

[0028] It should be noted that spatial distance The measurement is no longer based on the magnitude of torque, but rather on the similarity of the overall dynamic state of the system at two different moments. This is because the phase point simultaneously includes the current torque, historical torque, and travel speed. When the torque is very small, it means that the torque values ​​at these two moments are not only similar, but also that the recent evolution trend of the torque and the operating speed are similar, thus realizing the recursion of the system state recursion.

[0029] When the phase space trajectory undergoes continuous drift due to asymptotic congestion, the Euclidean distance between the new phase point and the historical phase point is... The system will increase systematically, causing many relationships that originally constituted recursion in the past to disappear, which is reflected in the recursive matrix as the diagonal structure becoming shorter and sparser.

[0030] The recursive feature construction module is used to determine whether two phase points constitute a recursive state based on the spatial distance between each phase point and a preset distance neighborhood threshold. The phase points that constitute a recursive state are counted as a recursive matrix with time as the row and column index.

[0031] The specific method for obtaining the preset distance neighborhood threshold is as follows: For any phase point in the current time window, determine the three phase points that are closest to it in space, and define the spatial distance between the phase point and the three closest phase points as three sets of nearest neighbor distances. Traverse all phase points in the current time window, summarize the three nearest neighbor distances of all phase points into a set, and use the product of the median of the set and the preset margin coefficient as the distance neighborhood threshold. It should be noted that the median has a natural robustness against outliers. Even if there are a few phase points in the window that are extremely far away from the main trajectory due to shocks, their nearest neighbor distance is extremely large, and they will only be placed at the end of the sorting sequence, without affecting the value of the median. Therefore, the median can stably reflect the typical neighborhood scale of the main phase point group. The purpose of the margin factor is to appropriately amplify the typical neighbor scale defined by the median, so that as long as the distance between two phase points does not exceed a certain multiple of the typical neighbor scale, they are considered close enough to be recursively detected. This multiple is based on empirical calibration and can better balance false positives and false negatives. The margin factor is generally set between 2 and 5.

[0032] The current time window is specifically defined as having the current time as the end point of the time window and determining the start point of the time window through a preset time width, thereby forming the current time window.

[0033] The specific method for constructing the recursive matrix is ​​as follows: The spatial distance between different phase points is compared with a distance neighborhood threshold. If the spatial distance between different phase points is not greater than the distance neighborhood threshold, then the two phase points constitute a recursive state. This is used to construct a binary recursive matrix, whose elements are defined as follows: In the formula, For the k-th row of the recursive matrix, the first... Column element values; It should be noted that by comparing the spatial distance between phase points with the adaptive threshold and then binarizing it to 0 or 1, a topological information space that retains only the adjacency relationship is constructed for the blockage perception of the chili harvester. In this scheme, the torque signal is reconstructed into an augmented phase space vector that integrates historical evolution and walking speed. The three physical processes of soil impact, progressive blockage, and steady-state fluctuation have completely different geometric distributions in this space: impact is manifested as an instantaneous outlier of a single phase point, with its distance from all phase points being extremely large; progressive blockage is manifested as the continuous drift of the phase trajectory, with new phase points gradually moving away from the historical phase point group; steady-state fluctuation is manifested as phase points repeatedly shuttling through a finite area. Through the construction of the recursive matrix, the impact outlier points naturally fall outside the threshold and do not constitute recursion, while the steady-state fluctuation points naturally fall within the threshold and constitute recursion. The binarized recursive matrix completely eliminates the influence of magnitude shift because the threshold itself adaptively floats with the local density of the phase point distribution. The threshold automatically shrinks when unloaded and automatically expands when heavily loaded, ensuring that the distribution density of 1 in the recursive matrix remains within a reasonable range that can reflect changes in the diagonal structure. This makes the diagonal weighted average length extracted in step three a universal indicator that is independent of the absolute magnitude of torque and only sensitive to the direction of dynamic drift. Regardless of the variety of chili peppers or the terrain of the field, the intensification of the blockage trend is uniformly manifested as a monotonically decreasing diagonal weighted average length, thus achieving true adaptive control of operating conditions.

[0034] The recursive feature judgment module is used to statistically analyze the distribution density of recursive points along the diagonal direction in the recursive matrix, extract the length distribution features of recursive segments along the diagonal direction, and calculate the recursive trend strength reflecting the mechanical stability of the harvesting equipment based on the length distribution features.

[0035] The logic behind extracting the length distribution features of consecutive recursive segments along the diagonal is as follows: Diagonal structure analysis is performed on the constructed recursive matrix, and from any recursive matrix... The process begins by scanning along the main diagonal, identifying segments with consecutive element values ​​of 1 along this direction. The length of each segment is recorded, with a minimum interval of 1 between adjacent element values. Segments with a length of at least 2 are designated as target segments. The frequency of each target segment is recorded, and the recursive trend strength is calculated based on this frequency, using the following formula: In the formula, For the strength of the recursive trend, The length of the target segment. The maximum length of the target segment. This indicates the frequency of occurrence of a target segment of length c.

[0036] It should be noted that the strength of the recursive trend As the core scalar index for measuring the dynamic stability of the drum-traveling coupling system of a chili harvester, when the chili harvester is in an ideal operating state of no-load uniform speed or uniform feeding, the drum torque fluctuates slightly around a certain mean, the traveling speed remains constant, and the phase space trajectory is confined to a compact low-dimensional attractor for repeated cycles. The phase points frequently recursively interact with each other, and the recursion matrix is ​​rich in a large number of long diagonal segments. A high value reflects that the system has strong dynamic inertia and high state repeatability, which means that the material flow inside the drum is in dynamic equilibrium and there is no continuous directional accumulation. When progressive blockage occurs, material accumulates continuously inside the drum, and the drum load torque shows a monotonically increasing trend. At the same time, the travel speed is passively reduced or actively reduced due to the increased load. The phase space trajectory no longer returns to the historical region, but continues to expand into the new state space. This drift causes the spatial distance between the newly generated phase point and the historical phase point to increase continuously, which cannot meet the recursion condition. The original long diagonal structure is cut into short segments or disappears completely, and the diagonal distribution in the recursion matrix is ​​concentrated at the low length end. The value has therefore continued to decline, and the rate and magnitude of this decline directly reflect the urgency of the congestion's development.

[0037] The high-torque pulse generated by instantaneous soil impact forms only isolated outlier phase points in phase space. These phase points may form sporadic recursions with other time points, but cannot form a continuous diagonal structure. Therefore, the weighted average length... His contribution was almost zero; this made The value is immune to occasional shocks and only responds to structural shifts in the system state, rather than transient disturbances.

[0038] A preset dynamic threshold value is used to determine the blockage status. This dynamic threshold value is specifically calculated dynamically based on historical data, while also incorporating a forgetting factor to ensure that the threshold value fluctuates with the normal operating conditions of the system. The specific formula used is as follows: In the formula, Let be the dynamic boundary value for the t-th time window. The recursive trend strength is the value of the (t-1)th time window. Forgetting factor, This is the dynamic boundary value for the (t-1)th time window, where t is the index of the time window; It should be noted that the formula uses an exponentially weighted moving average, making... Always tracking The recent level of the value becomes a dynamic estimate of the fluctuation range of the system under normal operating conditions; when When the value is lower than this dynamic estimate, it means that the degree of fragmentation of the current recursive structure has exceeded the reasonable range of the system's recent fluctuations, thus it can be inferred that there are additional trend factors, namely, gradual blockage or blockage is occurring. Forgetting factor This determines the response speed of the dynamic boundary value to recent observations. The larger, The more sensitive it is to the latest changes, the faster it can follow them, but the weaker its ability to resist interference. The smaller, The smoother and more stable the system, the more delayed its response to real-world operating conditions, and the greater the forgetting factor. The specific setting is based on expert experience, and the value is generally between 0.05 and 0.1.

[0039] The specific formula used to calculate the rate of change of the recursive trend strength is as follows: In the formula, Let be the rate of change of the recursive trend intensity in the t-th time window. Let be the recursive trend strength at the t-th time window; The current congestion status is determined based on the rate of change of recursive trend intensity combined with a dynamic boundary value. The specific logic is as follows: If both conditions are met Less than 0, and When the value is less than the current dynamic threshold, it is determined that the system is currently in a state of congestion. when Not less than 0 or If the value is not less than the current dynamic boundary value, it is determined that the current state is non-blocking.

[0040] It should be noted that when A value less than 0 indicates a decreasing recursive trend strength. A decreased recursive trend strength suggests reduced continuity of equipment operation, meaning an increased probability of congestion. This condition also ensures that the system has not reverted to a more stable direction. If the value is less than the current dynamic threshold, it means that the current recursive trend strength has fallen below the dynamic benchmark. This condition ensures that the current dynamic stability of the system is not only decreasing, but has also fallen below the lower limit of the recent normal fluctuation range. Therefore, it is judged to be in a blocked development state. Otherwise, it is in a non-blocked state or has been in a recovery state and no adjustment is required.

[0041] The speed control quantity generation module is used to compare the recursive trend intensity at the current moment with the recursive trend intensity at the previous sampling moment, calculate the rate of change, determine the type of blockage based on the rate of change and the current recursive trend intensity, and generate the speed reduction adjustment amount by combining the blockage type, the current recursive trend intensity, and the preset nonlinear compression function.

[0042] In non-blocking conditions, no adjustment of the operation speed is required; For the congestion development state, a nonlinear compression function is constructed. The nonlinear compression function is specifically set based on the intensity of the recursive trend within the time window, and the specific formula is as follows: In the formula, Let be the deceleration adjustment amount for the t-th time window. The maximum single speed adjustment range is set. This is the sensitivity coefficient; It should be noted that, Indicates the first The reduction adjustment amount within a time window is used to adjust the system's operating speed to alleviate congestion. The maximum single speed adjustment range represents the maximum speed reduction that the system can achieve within a time window; this value is set to limit the adjustment intensity and prevent the system from overreacting. in Used to determine the overall trend and magnitude of adjustments to the work speed, when The larger the value, the farther the system is from the critical state, that is, the more serious the current blockage. The sensitivity coefficient controls the sensitivity of the speed reduction adjustment to the current blockage state. The sensitivity coefficient is generally set based on historical experience and expert experience, and is usually set between 0.1 and 0.3. The speed adjustment process achieves non-linear change through an exponential function, which can accurately control the speed adjustment range and avoid oscillation or over-adjustment.

[0043] Based on the deceleration adjustment amount in the t-th time window, a first-order low-pass smoothing is applied to obtain the final target speed. The specific formula used to calculate the target speed is as follows: In the formula, Let be the target speed for the t-th time window. Let be the average actual walking speed in the t-th time window. Let the target velocity be the velocity in the (t-1)th time window. This is a smoothing coefficient, typically between 0.25 and 0.5; It should be noted that in practical applications, the walking speed and deceleration adjustment may contain noise or instantaneous fluctuations. This noise may cause rapid changes in the target speed, affecting system stability. Excessive target speed adjustment may cause the system to overreact. First-order low-pass smoothing can ensure that the system gradually adapts to changes and avoids the risks caused by sudden adjustments. This formula combines the previous target speed with the current actual speed and adjusts the target speed through a feedback mechanism, enabling the system to make intelligent adjustments based on historical and current states. The calculation method of the target speed in the formula is regarded as a weighted average of the historical target speed, the current actual speed, and the deceleration adjustment amount, which can better reflect the long-term trend of the system and is not disturbed by short-term fluctuations.

[0044] If the congestion type is a congestion development state, the speed reduction adjustment is the calculated value obtained by substituting the recursive trend strength into a preset nonlinear compression function and performing first-order low-pass smoothing. If the congestion type is a non-congestion state, the speed reduction adjustment is zero.

[0045] Please see Figure 2 The present invention also provides an adaptive control method for the operating speed of a chili harvesting device, wherein the adaptive control method for the operating speed of a chili harvesting device is used to control the aforementioned adaptive control system for the operating speed of a chili harvesting device, comprising: Step 1: Synchronously collect the instantaneous values ​​of the drum torque and travel speed of the harvesting equipment during the chili harvesting operation at fixed sampling intervals, and reconstruct the phase space of the instantaneous torque value sequence at multiple consecutive sampling times according to the preset time delay embedding parameters; Step 2: Combine the instantaneous torque value at any given moment with the instantaneous torque value of the preceding time delay to form a phase point corresponding to that moment. Incorporate the current walking speed value as an additional dimension into each phase point. Collect the phase points to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, calculate the spatial distance between it and other phase points in the trajectory. Step 3: Based on the spatial distance between each phase point, determine whether two phase points constitute a recursive state by using a preset distance neighborhood threshold, and count the phase points that constitute a recursive state into a recursive matrix with time as the row and column index. Step 4: Statistically analyze the distribution density of recursion points along the diagonal direction in the recursion matrix, extract the length distribution characteristics of recursion segments along the diagonal direction, and calculate the recursion trend strength reflecting the mechanical stability of the harvesting equipment based on the length distribution characteristics; Step 5: Compare the recursive trend strength at the current moment with the recursive trend strength at the previous sampling moment, calculate the rate of change, determine the type of blockage based on the rate of change and the current recursive trend strength, and generate the deceleration adjustment amount by combining the type of blockage, the current recursive trend strength, and the preset nonlinear compression function.

[0046] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0047] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0048] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0049] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An adaptive control system for the operating speed of a chili harvesting device, characterized in that, Specifically, it includes: The spatial phase point construction module is used to synchronously collect the instantaneous values ​​of the drum torque and the walking speed of the harvesting equipment during the chili harvesting operation at fixed sampling intervals, and to reconstruct the phase space of the torque instantaneous value sequence at multiple consecutive sampling times according to preset time delay embedding parameters. The phase trajectory generation module is used to combine the instantaneous torque value at any moment with the instantaneous torque value of the preceding time delay at that moment to jointly form the phase point corresponding to that moment. The current walking speed value is added as an additional dimension to each phase point, and the phase points are collected to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, the spatial distance between it and other phase points in the trajectory is calculated. The recursive feature construction module is used to determine whether two phase points constitute a recursive state based on the spatial distance between each phase point and a preset distance neighborhood threshold, and to count the phase points that constitute a recursive state into a recursive matrix with time as the row and column index. The recursive feature judgment module is used to statistically analyze the distribution density of recursive points along the diagonal direction in the recursive matrix, extract the length distribution features of recursive segments along the diagonal direction, and calculate the recursive trend strength reflecting the mechanical stability of the harvesting equipment based on the length distribution features. The speed control quantity generation module is used to compare the recursive trend intensity at the current moment with the recursive trend intensity at the previous sampling moment, calculate the rate of change, determine the type of blockage based on the rate of change and the current recursive trend intensity, and generate the speed reduction adjustment amount by combining the blockage type, the current recursive trend intensity, and the preset nonlinear compression function.

2. The adaptive control system for the operating speed of a chili harvesting device according to claim 1, characterized in that: The time delay embedding parameters specifically include time delay and phase point embedding dimension; the specific method for determining time delay and phase point embedding dimension is as follows: design field chili harvesting test, collect instantaneous torque value and instantaneous walking speed value under different harvesting conditions, the harvesting conditions include unloaded uniform speed, normal feeding, soil impact, gradual blockage and recovery after manual intervention to reduce speed; Based on the experimental roller torque time series data, the mutual information method is used to calculate the mutual information value between any roller torque time series data and the roller torque after the corresponding time delay. The specific steps are as follows: set several fixed time delays, which are set based on the sampling interval; calculate the mutual information value of the roller torque time series data under different fixed time delays; analyze the changes of all mutual information values ​​under each fixed time delay; and when the mutual information value first drops to a preset information threshold, use the corresponding fixed time delay as the time delay for setting the phase point. Given a fixed time delay, the phase point embedding dimension is set to increase incrementally. The pseudo-nearest neighbor method is used to calculate the pseudo-nearest neighbor ratio for different embedding dimensions. If the pseudo-nearest neighbor ratio drops to a preset ratio threshold or below for the first time, and the decrease in the pseudo-nearest neighbor ratio does not exceed a preset amplitude threshold when the embedding dimension is subsequently increased, then the embedding dimension is taken as the fixed embedding dimension.

3. The adaptive control system for the operating speed of a chili harvesting device according to claim 2, characterized in that: The phase point is constructed based on the time delay embedding parameters, specifically using the following formula: In the formula, Let k be the phase point at the k-th sampling time. Let be the instantaneous value of the torque at the k-th sampling time. For the k-th The instantaneous value of torque at the sampling time. For a fixed time delay, To fix the embedding dimension, Let k be the instantaneous walking speed at the k-th sampling time; k is the index of the sampling time. For any two phase points in the augmented phase space trajectory, their Euclidean distance is calculated as the spatial distance. The specific logic behind this calculation is as follows: In the formula, Specifically, it represents the k-th phase point and the... Spatial distance between phase points Represents the kth phase point. One portion, Indicates the first The th phase point Each component, k and The index of the phase point, and .

4. The adaptive control system for the operating speed of a chili harvesting device according to claim 1, characterized in that, The specific method for obtaining the preset distance neighborhood threshold is as follows: For any phase point within the current time window, determine the three phase points that are closest to it in space, and define the spatial distance between the phase point and the three closest phase points as three sets of nearest neighbor distances. Traverse all phase points within the current time window, summarize the three nearest neighbor distances of all phase points into a set, and use the product of the median of the set and the preset margin coefficient as the distance neighborhood threshold. The current time window is specifically defined as having the current time as the end point of the time window and determining the start point of the time window through a preset time width, thereby forming the current time window.

5. The adaptive control system for the operating speed of a chili harvesting device according to claim 4, characterized in that, The specific method for constructing the recursive matrix is ​​as follows: The spatial distance between different phase points is compared with a distance neighborhood threshold. If the spatial distance between different phase points is not greater than the distance neighborhood threshold, then the two phase points constitute a recursive state. This is used to construct a binary recursive matrix, whose elements are defined as follows: In the formula, For the k-th row of the recursive matrix, the first... Column element values; The logic behind extracting the length distribution features of consecutive recursive segments along the diagonal is as follows: Diagonal structure analysis is performed on the constructed recursive matrix, and from any recursive matrix... The process begins by scanning along the main diagonal, identifying segments with consecutive element values ​​of 1 along this direction. The length of each segment is recorded, with a minimum interval of 1 between adjacent element values. Segments with a length of at least 2 are designated as target segments. The frequency of each target segment is recorded, and the recursive trend strength is calculated based on this frequency, using the following formula: In the formula, For the strength of the recursive trend, The length of the target segment. The maximum length of the target segment. This indicates the frequency of occurrence of a target segment of length c.

6. The adaptive control system for the operating speed of a chili harvesting device according to claim 1, characterized in that: A preset dynamic threshold value is used to determine the blockage status. This dynamic threshold value is specifically calculated dynamically based on historical data, while also incorporating a forgetting factor to ensure that the threshold value fluctuates with the normal operating conditions of the system. The specific formula used is as follows: In the formula, Let be the dynamic boundary value for the t-th time window. The recursive trend strength is at the (t-1)th time window. Forgetting factor, This is the dynamic boundary value for the (t-1)th time window, where t is the index of the time window; The specific formula used to calculate the rate of change of the recursive trend strength is as follows: In the formula, Let be the rate of change of the recursive trend intensity in the t-th time window. Let be the recursive trend strength at the t-th time window; The current congestion status is determined based on the rate of change of recursive trend intensity combined with a dynamic boundary value. The specific logic is as follows: If both conditions are met Less than 0, and When the value is less than the current dynamic threshold, it is determined that the system is currently in a state of congestion. when Not less than 0 or If the value is not less than the current dynamic boundary value, it is determined that the current state is non-blocking.

7. The adaptive control system for the operating speed of a chili harvesting device according to claim 6, characterized in that: In non-blocking conditions, no adjustment of the operation speed is required; For the congestion development state, a nonlinear compression function is constructed. The nonlinear compression function is specifically set based on the intensity of the recursive trend within the time window, and the specific formula is as follows: In the formula, Let be the deceleration adjustment amount for the t-th time window. The maximum single speed adjustment range is set. This is the sensitivity coefficient; Based on the deceleration adjustment amount in the t-th time window, a first-order low-pass smoothing is applied to obtain the final target speed. The specific formula used to calculate the target speed is as follows: In the formula, Let be the target speed for the t-th time window. Let be the average actual walking speed in the t-th time window. Let the target velocity be for the (t-1)th time window. For smoothing coefficients; If the congestion type is a congestion development state, the speed reduction adjustment is the calculated value obtained by substituting the recursive trend strength into a preset nonlinear compression function and performing first-order low-pass smoothing. If the congestion type is a non-congestion state, the speed reduction adjustment is zero.

8. A method for adaptive control of the operating speed of a chili harvesting device, used to control the adaptive control system for the operating speed of a chili harvesting device as described in any one of claims 1-7, characterized in that, include: During chili harvesting operations, the instantaneous values ​​of the drum torque and travel speed of the harvesting equipment are collected synchronously at fixed sampling intervals. The sequence of instantaneous torque values ​​from multiple consecutive sampling times is then reconstructed in phase space according to preset time delay embedding parameters. The instantaneous torque value at any given moment is combined with the instantaneous torque value of the preceding time delay to form a phase point corresponding to that moment. The current walking speed value is added as an additional dimension to each phase point. The phase points are aggregated to form an augmented phase space trajectory. For each phase point in the augmented phase space trajectory, the spatial distance between it and other phase points in the trajectory is calculated. Based on the spatial distance between each phase point, a preset distance neighborhood threshold is used to determine whether two phase points constitute a recursive state. The phase points that constitute a recursive state are counted as a recursive matrix with time as the row and column index. The distribution density of recursion points along the diagonal direction in the recursion matrix is ​​statistically analyzed, the length distribution characteristics of recursion segments along the diagonal direction are extracted, and the recursion trend strength reflecting the mechanical stability of the harvesting equipment is calculated based on the length distribution characteristics. The recursive trend strength at the current moment is compared with the recursive trend strength at the previous sampling moment to obtain the rate of change. The type of congestion is determined based on the rate of change and the current recursive trend strength. The deceleration adjustment amount is generated by combining the congestion type, the current recursive trend strength, and the preset nonlinear compression function.

Images